From: Qu Wenruo Date: Thu, 3 Apr 2025 23:40:51 +0000 (+1030) Subject: btrfs: use folio_contains() for EOF detection X-Git-Tag: block-6.16-20250606~42^2~151 X-Git-Url: https://git.kernel.dk/?a=commitdiff_plain;h=2b14b74b992187933069cb3316c691dbd5707f47;p=linux-block.git btrfs: use folio_contains() for EOF detection Currently we use the following pattern to detect if the folio contains the end of a file: if (folio->index == end_index) folio_zero_range(); But that only works if the folio is page sized. For the following case, it will not work and leave the range beyond EOF uninitialized: The page size is 4K, and the fs block size is also 4K. 16K 20K 24K | | | | | EOF at 22K And we have a large folio sized 8K at file offset 16K. In that case, the old "folio->index == end_index" will not work, thus the range [22K, 24K) will not be zeroed out. Fix the following call sites which use the above pattern: - add_ra_bio_pages() - extent_writepage() Reviewed-by: Filipe Manana Signed-off-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba --- diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 56d970434b4b..74edd2dd92ff 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -523,7 +523,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, free_extent_map(em); unlock_extent(tree, cur, page_end, NULL); - if (folio->index == end_index) { + if (folio_contains(folio, end_index)) { size_t zero_offset = offset_in_folio(folio, isize); if (zero_offset) { diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 12a668dd5e5e..ba067d3461e6 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -221,7 +221,7 @@ static void __process_folios_contig(struct address_space *mapping, } static noinline void unlock_delalloc_folio(const struct inode *inode, - const struct folio *locked_folio, + struct folio *locked_folio, u64 start, u64 end) { ASSERT(locked_folio); @@ -231,7 +231,7 @@ static noinline void unlock_delalloc_folio(const struct inode *inode, } static noinline int lock_delalloc_folios(struct inode *inode, - const struct folio *locked_folio, + struct folio *locked_folio, u64 start, u64 end) { struct btrfs_fs_info *fs_info = inode_to_fs_info(inode); @@ -1707,7 +1707,7 @@ static int extent_writepage(struct folio *folio, struct btrfs_bio_ctrl *bio_ctrl return 0; } - if (folio->index == end_index) + if (folio_contains(folio, end_index)) folio_zero_range(folio, pg_offset, folio_size(folio) - pg_offset); /*